Gemini Image Proxy

Generate and edit images with Gemini API using the OpenAI Python SDK.

Security Scan

The skill's code, instructions, and environment needs match its stated purpose (generating/editing images via an OpenAI-compatible Gemini image endpoint); nothing appears incoherent or excessive, but you should only provide a trusted base URL and API key because prompts and images are sent to that endpoint.

Purpose & Capability

Name/description say: generate/edit images via Gemini using the OpenAI Python SDK. The skill requires an API key and base URL for an OpenAI-compatible endpoint and installs/uses the openai SDK — all are appropriate and expected.

Instruction Scope

SKILL.md and scripts/generate.py limit actions to calling client.images.generate and client.images.edits, reading an optional local input image, and writing an output image. This is within scope, but the skill will transmit prompts and any input image to the configured base URL — verify that endpoint is trusted.

Install Mechanism

No packaged install spec; instructions ask to pip install the openai package only. No downloads from arbitrary URLs or extraction steps are present.

Credentials

The two required environment variables (GOOGLE_PROXY_API_KEY and GOOGLE_PROXY_BASE_URL) are proportional to the task. However, the provided API key grants whatever access the configured base URL accepts, so only supply credentials for endpoints you trust.

Persistence & Privilege

always is false and the skill does not request system-wide changes or persistent privileges. It does not modify other skills or store global config.

Assessment

This skill appears to do what it claims: call an OpenAI-compatible image generation/edit API and save images locally. Before installing, ensure GOOGLE_PROXY_BASE_URL points to a trusted service (the skill will send your prompts and any input images there) and that you are comfortable with the API key's permissions. Installing the openai Python package is required. If you do not trust the base URL or key, do not provide them — the skill has no other hidden capabilities.

Gemini Image Simple

Generate and edit images using Gemini 3 Pro Image via the OpenAI Python SDK and an OpenAI-compatible API endpoint.

Why This Skill

Feature	This Skill	Others (nano-banana-pro, etc.)
Dependencies	openai (SDK)	google-genai, pillow, etc.
Requires pip/uv	✅ Yes	✅ Yes
Works on Fly.io free	✅ Yes (with pip)	❌ Fails
Works in containers	✅ Yes (with pip)	❌ Often fails
Image generation	✅ Full	✅ Full
Image editing	✅ Yes	✅ Yes
Setup complexity	Install SDK + set API key	Install packages first

Bottom line: This skill uses the OpenAI SDK, so you must install openai once with pip.

Install

python3 -m pip install openai

Quick Start

# Set env
export GOOGLE_PROXY_API_KEY="your_api_key"
export GOOGLE_PROXY_BASE_URL="https://example.com/v1"

# Generate
python3 /data/clawd/skills/gemini-image-simple/scripts/generate.py "A cat wearing a tiny hat" cat.png

# Edit existing image
python3 /data/clawd/skills/gemini-image-simple/scripts/generate.py "Make it sunset lighting" edited.png --input original.png

Usage

Generate new image

python3 {baseDir}/scripts/generate.py "your prompt" output.png

Edit existing image

python3 {baseDir}/scripts/generate.py "edit instructions" output.png --input source.png

Supported input formats: PNG, JPG, JPEG, GIF, WEBP

Environment

Set these environment variables:

• GOOGLE_PROXY_API_KEY (your API key)
• GOOGLE_PROXY_BASE_URL (OpenAI-compatible base URL, e.g. https://example.com/v1)

How It Works

Uses Gemini 3 Pro Image (gemini-3-pro-image) via the OpenAI Python SDK:

• client.images.generate(...) for new images
• client.images.edits(...) for edits
• Requires the openai package

That's it. Works on any Python 3.10+ installation with openai installed.

Model

Currently using: gemini-3-pro-image

Other available models (can be changed in generate.py if needed):

• gemini-3-pro-image-preview - Preview variant
• imagen-4.0-ultra-generate-001 - Imagen 4.0 Ultra
• imagen-4.0-generate-001 - Imagen 4.0
• gemini-2.5-flash-image - Gemini 2.5 Flash with image gen

Examples

# Landscape
python3 {baseDir}/scripts/generate.py "Misty mountains at sunrise, photorealistic" mountains.png

# Product shot
python3 {baseDir}/scripts/generate.py "Minimalist product photo of a coffee cup, white background" coffee.png

# Edit: change style
python3 {baseDir}/scripts/generate.py "Convert to watercolor painting style" watercolor.png --input photo.jpg

# Edit: add element
python3 {baseDir}/scripts/generate.py "Add a rainbow in the sky" rainbow.png --input landscape.png